Key Points

Introduction to OpenRefine

  • OpenRefine is a free, open-source tool for cleaning, organizing, and exploring messy data.
  • You can easily import, filter, sort, and analyze your data, even without technical experience.
  • OpenRefine supports many data formats and can be extended with add-ons and custom scripts for even more possibilities.
  • Using OpenRefine helps you prepare your data for analysis, supporting transparent and reproducible research practices.

Importing Data and Getting to Know the OpenRefine User Interface

  • OpenRefine organizes your work in projects
  • You can import data from different sources and in different formats into OpenRefine
  • Adjust import settings to ensure your data is read correctly and preview the results before starting
  • The main components of the user interface are the grid, the grid header, the project bar, and Facet/Filter as well as Undo/Redo tab

Exploring Data

  • Facets provide an interactive overview of the values in a column and help you explore your data.
  • Multi-valued cells must be split before accurate faceting is possible.
  • Numeric and Timeline facets require converting text values into numbers or dates first.

Custom Facets and GREL

  • Custom facets group data using computed results from a GREL expression, not only the original cell values.
  • GREL is a lightweight language that allows you to inspect, classify, and analyse data inside OpenRefine.
  • Custom facets let you ask flexible questions about your data, such as identifying multiple creators or unusually long titles.
  • With conditional expressions like if(), you can define new categories that support deeper exploration and data-quality checks.
  • GREL functions can be chained together to answer more complex questions about your data.

Transforming Data

  • GREL expressions can be used to extract, modify, and standardize information.
  • Creating new columns preserves the original data and makes transformations easier to review.
  • The split() function creates arrays that can be accessed using positions such as [0] and [1].
  • Structured information can be extracted from text using GREL expressions and pattern matching.
  • Data cleaning often requires multiple transformation steps and manual review.
  • Clustering helps identify potentially equivalent values and supports manual standardization.

Reconciling Data with External Data Sources

  • Reconciliation links text strings to unique identifiers in external databases.
  • This makes your dataset more precise, reusable, and comparable across projects.
  • OpenRefine suggests matches, but users should always review and confirm them.
  • Identifier columns preserve these links when exporting the dataset.

Exporting and Importing Data and Workflows

  • OpenRefine records every transformation you make.
  • The Undo/Redo tab lets you move backward and forward through your cleaning process.
  • Workflows can be exported as JSON and reapplied to other projects, ensuring transparency and reproducibility.